Planner for Relevant Policies
Disclaimer: PRP is going through some major changes at the moment. Until the new version is fully implemented and the associated publication submitted, it is recommended that you use [this version] of the PRP code which contains many bug fixes and performance improvements over this repository. If you want to use PRP in any published work, please consider contacting me for a version that is most appropriate. See also here for details on comparing with PRP.
Welcome to the page for PRP -- a state-of-the-art Fully Observable Non-Deterministic (FOND) planner.
- Usage: Information on the command-line options for PRP.
- Changelog: A log of the changes PRP has gone through.
- Script Usage: Information on the scripts included with PRP.
- Comparing with PRP: Information on comparing with PRP.
Variants of PRP
- Cond-PRP: PRP now supports conditional effects natively.
- PO-PRP: Version capable of solving planning problems with partial observability and sensing.
- MA-PRP: Version capable of solving first-person multi-agent planning problems.
- Prob-PRP: Version capable of solving probabilistic planning problems (i.e., PPDDL).
Papers / Presentations
- (ICAPS'12) Improved Non-deterministic Planning by Exploiting State Relevance
- (ICAPS'14) Non-Deterministic Planning With Conditional Effects (slides) (poster)
- (AAAI'14) Computing Contingent Plans via Fully Observable Non-Deterministic Planning (slides)
- (HSDIP'15) From FOND to Probabilistic Planning: Guiding search for quality policies (slides) (poster)
- (DMAP'15) Leveraging FOND Planning Technology to Solve Multi-Agent Planning Problems (slides) (poster)